Flex Fuel Issues...

[AbeFM]

Er, so the other day I noticed that at higher speeds my car would have some major stumbling. Sounded like it was missing.

I looked at the logs, so no knock retard, nice AFR's, stable fuel pressure, everything looks fine but the car was NOT right - after a second or so I pulled out. It happened a couple more times.

In the logs, I see both of the Flex Fuel sensor readings going all to heck! From a relatively steady 8.2% (typically I see 8-8.5) and a temp of 96 degrees to 90% and -40.

While the temp going to minimum makes sense, why the temp would not only go to a weird value, but it asymptotically approached normal again after the 1 second event....

Perhaps it's just losing power or ground to the sensor, but I wanted to know if this is something others are seeing - or anything else I should look into. Find logs and MSQ attached.
-Abe.

P.S. File is joined at 13345s - between 2nd and 3rd events just to show more of them.

[elutionsdesign]

Where is the sensor plumbed in to the fuel line?

[AbeFM]

On the return from the fuel pressure regulator which itself is after the outlet of the fuel rail. It's under the hood.

[AbeFM]

My two thoughts on the issue (well, three really) are:

1) Bad wiring
2) Some kinda code issue where high counts on the VSS interfere - I've only seen the issue over ~85 mph, and I've seen it most of the times I've been there.
3) FPR not feeding enough fuel to the sensor to keep it working.

1 is reasonable, but the other sensors (like the crank and cam position) didn't get weird. And the battery voltage are nominal - so it'd have to be wiring directly to the sensor... And why doesn't it happen at other periods of high load?

2 seems unlikely, but it is something I know too little about to say yea or nay.

3 also seems unlikely. The fuel pressure doesn't fall off, and the pump supposedly can supply 400 lph, I'm sure I'm not touching half of that. Again, I don't see this at high engine speed or load, just high road speed.

Anyway, for right now I just turned off the corrections, but it's not much of a solution and any help is appreciated. Hopefully I'll be able to log more events in the future in other conditions to illuminate things more. Perhaps first I'll just wiggle the wires. (EDIT: Extreme wire wiggling didn't change anything, and there's still no evidence of fuel leaks. Sensor itself could be bad, but I would expect it to go bad more randomly)

[elutionsdesign]

Personally I'd like to see some extreme filtering on the flex fuel sensors, they can't possibly need the constant updating required of the coolant or manifold air temperature sensors. Having said that you should try the sensor in the feed side of the rail instead of the return.

[AbeFM]

I agree with the filtering suggestion - and they don't need to be read often either. But in this case, that would just disguise the issue and that is a concern.

They specifically call out NOT to put them on the upstream side. They are rated for 400 lph, but for low pressure only. There is so much misinformation - I've seen several online "pin outs" which don't match what's stamped right on the device.

[jsmcortina]

This looks like a signal error or a signal processing error in the code.

I checked and there was a code change relating to the flex fuel temperature calculation in the pre-1.4 code. I'm not seeing anything else at the moment.

I'd like to get you some debug 1.4 code to record the raw data before the flex fuel calcs to see if we can establish what is going wrong.

James

[AbeFM]

Sounds like a great idea. I'm on the release code now. Shot me over whatever, pm or email is fine.

Any ideas or just kinda looking hoping to see something?

Could those asymptotic returns be averaging in the ms?

[jsmcortina]

The firmware already had code to reject rogue pulses, however I found a few mistakes in it that were preventing it from working.

I've sent a test/debug code by email. Post up how you get on.

James

[ttxpc]

Is this an install specific issue or general issue with the firmware/code?

Reason I ask is I'm about to install a flex fuel sensor this coming weekend, is this issue likely to affect my setup?

Cheers

[AbeFM]

You can do what I did - just set the scale flat (fuel 100% at 0% eth, 100% at 100%eth), and similar for temperature. Then it's physically in place and being logged. If it turns out to be software, it'll get fixed then you can turn it back on.

I've loaded the testing firmware, planning on driving some today to check it out.

[gui67]

Hello James,

could you please also send me the test code, I have the same behaviour with spikes in Ethanol percentage. I could test it on Friday.
The problem seems to appear mostly in high load situations, I don't know why. I plan to change the sensor and improve wiring this winter.

Thanks
Guillaume

[AbeFM]

that's interesting, and good to have a confirmation. I wouldn't say high load, precisely. I've seen it at high speed (oh, and the same "return from 0" behavior when I killed the motor mid-log and restarted it)... But haven't since I got the test code. Then again, I've got like <4 minutes on the clock. But I DO have yogurt.
-Abe.

[jsmcortina]

Hello James,

could you please also send me the test code, I have the same behaviour with spikes in Ethanol percentage. I could test it on Friday.
The problem seems to appear mostly in high load situations, I don't know why. I plan to change the sensor and improve wiring this winter.

Thanks
Guillaume

PM me your email and let me know MS3 or MS3-Pro, then I'll sent it over.

James

[AbeFM]

Looking at my logs from last night - where I didn't go over 88 mph - I DO get some tendencies for the fuel pressure and content to jump and slowly recover... But they are SMALL amounts - from ~8 - ~7 (8.2 to 6.9 was the biggest swing), and other than perhaps a correlation to flow I don't see anything terribly odd.

I'll stick with it a bit more - but it's looking more and more like just taking a 1 minute (or even 10 second) average would fix issues like these.
-Abe.

[jsmcortina]

But I DO have yogurt.

Any more testing updates?

James

[AbeFM]

I sent you the log I took, and will try to get another today, it's tough to find good opportunities to test at 100 mph.

[gui67]

Hello James,
I tested the code you sent.
I did not get any spike during the _minute drive, the Ethanol %age stayed between 72,5 and 73,4. with the earlier codes, I had at least 1 spike per minute, so for me, the problem seems solved. I will make a longer log tomorrow.

Thanks
Guillaume

[jsmcortina]

OK, good. Just to note that the version I sent you had additional changes to the one Abe received.

James

[gui67]

About the additionnal Flex Blend you mentionned in your email, does it have more capabilities then the VE, Spark, AFR, Boost, Crank% of the older versions? I see alternate Overboost and accel enrichment
Is it possible to blend idle VE?